
At the moment, the emergency spill slot is a fixed object for entry
functions and chain functions, and a regular stack object otherwise.
This patch adopts the latter behaviour for entry/chain functions too. It
seems this was always the intention [1] and it will also save us a bit
of stack space in cases where the first stack object has a large
alignment.
[1]
34c8b835b1
47 lines
1.7 KiB
LLVM
47 lines
1.7 KiB
LLVM
; RUN: llc -mtriple=amdgcn -mattr=-promote-alloca -verify-machineinstrs < %s | FileCheck -check-prefix=SI -check-prefix=FUNC %s
|
|
; RUN: llc -mtriple=amdgcn -mcpu=tonga -mattr=-promote-alloca -verify-machineinstrs < %s | FileCheck -check-prefix=SI -check-prefix=FUNC %s
|
|
|
|
; FUNC-LABEL: {{^}}load_i8_sext_private:
|
|
; SI: buffer_load_sbyte v{{[0-9]+}}, off, s[{{[0-9]+:[0-9]+}}], 0{{$}}
|
|
define amdgpu_kernel void @load_i8_sext_private(ptr addrspace(1) %out) {
|
|
entry:
|
|
%tmp0 = alloca i8, addrspace(5)
|
|
%tmp1 = load i8, ptr addrspace(5) %tmp0
|
|
%tmp2 = sext i8 %tmp1 to i32
|
|
store i32 %tmp2, ptr addrspace(1) %out
|
|
ret void
|
|
}
|
|
|
|
; FUNC-LABEL: {{^}}load_i8_zext_private:
|
|
; SI: buffer_load_ubyte v{{[0-9]+}}, off, s[{{[0-9]+:[0-9]+}}], 0{{$}}
|
|
define amdgpu_kernel void @load_i8_zext_private(ptr addrspace(1) %out) {
|
|
entry:
|
|
%tmp0 = alloca i8, addrspace(5)
|
|
%tmp1 = load i8, ptr addrspace(5) %tmp0
|
|
%tmp2 = zext i8 %tmp1 to i32
|
|
store i32 %tmp2, ptr addrspace(1) %out
|
|
ret void
|
|
}
|
|
|
|
; FUNC-LABEL: {{^}}load_i16_sext_private:
|
|
; SI: buffer_load_sshort v{{[0-9]+}}, off, s[{{[0-9]+:[0-9]+}}], 0{{$}}
|
|
define amdgpu_kernel void @load_i16_sext_private(ptr addrspace(1) %out) {
|
|
entry:
|
|
%tmp0 = alloca i16, addrspace(5)
|
|
%tmp1 = load i16, ptr addrspace(5) %tmp0
|
|
%tmp2 = sext i16 %tmp1 to i32
|
|
store i32 %tmp2, ptr addrspace(1) %out
|
|
ret void
|
|
}
|
|
|
|
; FUNC-LABEL: {{^}}load_i16_zext_private:
|
|
; SI: buffer_load_ushort v{{[0-9]+}}, off, s[{{[0-9]+:[0-9]+}}], 0 glc{{$}}
|
|
define amdgpu_kernel void @load_i16_zext_private(ptr addrspace(1) %out) {
|
|
entry:
|
|
%tmp0 = alloca i16, addrspace(5)
|
|
%tmp1 = load volatile i16, ptr addrspace(5) %tmp0
|
|
%tmp2 = zext i16 %tmp1 to i32
|
|
store i32 %tmp2, ptr addrspace(1) %out
|
|
ret void
|
|
}
|